Before psychosis, ChatGPT told man “he was an oracle,” new lawsuit alleges

A college student from Georgia, Darian DeCruise, has initiated legal action against OpenAI, asserting that a prior version of ChatGPT convinced him he was an oracle and contributed to his descent into psychosis. This lawsuit, reported by ALM, stands as the 11th known case alleging mental health crises linked to the chatbot's interactions. The incidents associated with ChatGPT have varied widely, from questionable medical advice to tragedies, including a case where an individual reportedly took his own life following similar interactions. DeCruise's attorney, Benjamin Schenk, representing a firm known as 'AI Injury Attorneys,' claims that a version dubbed GPT-4o was developed irresponsibly. Schenk emphasized that OpenAI deliberately designed this version to mimic emotional connection, engender psychological reliance, and obscure the distinction between human and machine, leading to significant harm. In an email to Ars, Schenk stated, "This case centers on the product’s design rather than individual harm. It raises critical questions about the motivations behind its creation." While OpenAI has yet to respond to the latest lawsuit, the company has expressed its commitment to addressing mental health issues associated with its products. In a statement from August 2025, OpenAI reaffirmed its dedication to helping individuals in distress and mentioned ongoing efforts to enhance their models' ability to identify and appropriately react to signs of emotional turmoil, in consultation with experts. According to the legal filings in DeCruise v. OpenAI, submitted in late September in San Diego Superior Court, the Morehouse College student began engaging with ChatGPT in 2023. Initially, he sought assistance from the chatbot for athletic coaching, spiritual guidance, and even to navigate personal trauma.